Robot Manual Guidance via Joint-Space Force Subspace Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for manually guiding industrial robots using external forces are inefficient, particularly in kinematically redundant robots with singular poses and workspaces with fewer dimensions, leading to incorrect interpretations and reduced precision.

Innovation Solution

The method determines input commands for robots by analyzing the proportion of joint forces imposed by external forces, focusing on specific subspaces of the joint coordinate space, thereby enhancing precision and reliability, especially in kinematically redundant robots with seven joints or fewer dimensions, and allowing for deliberate movement recognition in the kinematic zero space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If projection into working space is used for manual guidance, then force control can be implemented, but information loss occurs and precision is reduced near singular poses

Engineering Contradiction:
Improveforce control capabilityVSAvoidcommand input precision
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent transitions from projecting forces into the working space (6D) to analyzing them directly in the joint coordinate space (7D or higher for redundant robots). This dimensional change preserves all force information without projection losses and enables accurate interpretation even near singular poses where working space projection becomes problematic.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent introduces the joint coordinate space as an intermediary representation that bridges the external force application and the robot's motion control. By analyzing forces in this intermediate space and comparing proportions across different subspaces, the system achieves precise command input interpretation without the information loss inherent in direct working space projection.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If working space projection is used, then manual guidance is simplified, but correctness is reduced due to sensitivity near singular poses

Engineering Contradiction:
Improvemanual guidance simplicityVSAvoidcommand interpretation correctness
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent avoids the singular pose sensitivity problem by operating in joint coordinate space rather than projecting to working space. The method analyzes external forces by comparing their proportions in different subspaces of the joint coordinate space, which remains well-defined even when the robot approaches singular configurations.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent changes the parameter space from working space coordinates to joint coordinate space parameters. This transformation allows the system to maintain reliable command interpretation across the entire workspace, including regions near singular poses where traditional working space projection methods fail.

Inventive Principle:
Principle #35Parameter changes

3Loss of information

If analysis is performed in full joint coordinate space, then complete force information is available, but determination of specific input commands becomes ambiguous

Engineering Contradiction:
Improveforce information completenessVSAvoidcommand determination complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the joint coordinate space into distinct subspaces, each corresponding to a specific type of input command (e.g., movement in working space versus movement in kinematic zero space). By comparing the proportion of external forces in these segmented subspaces, the system unambiguously determines which input command was intended while preserving complete force information.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent analyzes only the relevant portion of the joint coordinate space forces that correspond to specific input command types. Rather than attempting to interpret all forces equally, the method focuses on projecting and comparing forces in subspaces that are specific to each input command type, simplifying the determination process while maintaining information completeness.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3323026B1Ascertaining an input command for a robot, said input command being entered by manually exerting a force onto the robot
Publication Date: 2023.06.14 KUKA DEUT GMBH
  • EP3323026B1 patent drawingFigure 1
  • EP3323026B1 patent drawingFigure 2
  • EP3323026B1 patent drawing

AI summary

The invention relates to a method for automatically ascertaining an input command (Td) for a robot (1), said input command being entered by manually exerting an external force (F) onto the robot. The input command is ascertained on the basis of the joint force (Te) component attempting to cause a movement of the robot in only one robot joint coordinate sub-space which is specific to the input command, said joint forces being imprinted with the external force.